ICSH Recommendations for the Measurement of Haemoglobin A2 A

International Journal of Laboratory Hematology The Ofﬁcial journal of the International Society for Laboratory Hematology

ORIGINAL ARTICLE INTERNATIONAL JOURNAL OF LABORATORY HEMATOLOGY

ICSH recommendations for the measurement of Haemoglobin A2 A. D. STEPHENS*, M. ANGASTINIOTIS†,E.BAYSAL‡,V.CHAN§,S.FUCHAROEN–,P.C.GIORDANO**, J. D. HOYER††,A.MOSCA‡‡,B.WILD§§, ON BEHALF OF THE INTERNATIONAL COUNCIL FOR THE STANDARDISATION OF HAEMATOLOGY (ICSH)

*Department of Haematology, SUMMARY University College London Hospitals, London, UK Although DNA analysis is needed for characterization of the mutations †Nicosia, Cyprus ‡Genetic and Thalassaemia Unit, Dubai that cause b-thalassaemia, measurement of the Hb A2 is essential for Health Authority, UAE the routine identification of people who are carriers of b-thalassaemia. § Department of Medicine, The The methods of quantitating Hb A are described together with pitfalls University of Hong Kong, Queen Mary 2 Hospital, Pokfulam, Hong Kong in undertaking these laboratory tests with particular emphasis on –Thalassemia Research Center, Institute automated high-performance liquid chromatography and capillary of Molecular Biosciences, Mahidol University, Thailand electrophoresis. **Hemoglobinopathies Reference Laboratory, Human and Clinical Genetics Department, Leiden University Medical Center, Einthovenweg, Leiden, The Netherlands ††Division of Hematopathology, The Mayo Clinic College of Medicine, Rochester, MN, USA ‡‡Dip. di Scienze e Tecnologie Biomediche, Universita` degli Studi di Milano, Milano, Italy §§UK NEQAS(H), Watford, UK

Correspondence: Dr A. D. Stephens, Department of Haematology, University College London Hospitals, 250 Euston Road, London, NW1 2PJ, UK. Tel.: 02034479638; Fax: 02034479911; E-mail: [email protected] doi:10.1111/j.1751-553X.2011.01368.x

Received 13 April 2011; accepted for publication 13 June 2011

Keywords Hb A2, HPLC, capillary electrophoresis, recommended methods, haemoglobinopathy, thalassaemia

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 1 2 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION

trait, ‘normal-Hb A ’ b-thalassaemia, db-thalassaemia INTRODUCTION 2 deletions, or a combination of these with common, or It is now 30 years since the ﬁrst recommendations uncommon, d-thalassaemia or d-chain variants or were published concerning the quantiﬁcation of more complex conditions that may interfere with Hb

Haemoglobin A2, and during this time, there have A2 measurement (Bouva et al., 2006; Phylipsen et al., been several new analytical developments in the 2011). It is well known that iron deficiency anaemia field; therefore, the International Council for the results in a wide range of red cell anomalies such as Standardisation of Haematology (ICSH) Board con- reduction in MCV, MCH, RBC and Hb, depending on sider that the original recommendations should be the severity at the time of analysis, whereas in carriers revised. Haemoglobin A2 (Hb A2) is a haemoglobin of thalassaemia, MCH, MCV and the Hb are reduced tetramer composed of two a and two d-globin but the red cell count is relatively high compared with chains (a2d2). It is a normal haemoglobin that is the haemoglobin level. Iron deficiency anaemia may not present in the fetus but gradually develops after be confused with a-thalassaemia or b-thalassaemia, birth increasing in concentration as d-chain produc- and therefore, it is helpful to undertake iron status tion occurs. The concentration of Hb A2 (the minor studies (such as serum ferritin, serum iron with trans- adult haemoglobin) is raised in most carriers of ferrin saturation and/or zinc protoporphyrin). In some b-thalassaemia, and the correct quantification of Hb cases, ‘high Hb A2 b-thalassaemia’ may occur together

A2 is essential for the routine diagnosis of b-thalas- with iron deficiency anaemia, and in most of these saemia trait. As b-thalassaemia major has major cases, the level of Hb A2 will remain elevated but in clinical implications for affected individuals and their some people with ‘borderline Hb A2 beta-thalassae- families, it is important to detect the presence of mia’, a significant iron depletion may reduce the b-thalassaemia trait in prospective parents so as to slightly elevated Hb A2 level into the normal range. In identify couples at risk of having a child with such situations, it would be advisable to treat the b-thalassaemia major. In countries where b-thalas- patient with iron to correct anaemia before determin- saemia is a significant national health problem, such ing Hb A2 levels. This is particularly important in as Cyprus, premarital screening tests have been rec- regions where both iron deficiency anaemia and a- ommended since the 1970s. In some countries such thalassaemia coexist in epidemic proportions such as as Iran, UAE and Bahrain, similar screening has in the United Arab Emirates where the frequency of only recently been introduced. In some countries the a-thalassaemia gene among the national popula- (such as England), such a test has to be offered to tion is estimated at around 50% through DNA studies all women in the early stages of pregnancy, but (Baysal, 2001). they are free to decline this testing if they so wish. In the diagnosis of b-thalassaemia trait, it is the pro-

In many countries in SE Asia where thalassaemia is portion of Hb A2 relative to any other haemoglobin prevalent, upon booking at the antenatal clinic, all present, not the absolute amount of Hb A2, that is clini- pregnant women with mean cell volume (MCV) cally important, and if a Hb A2 variant is present <80 fl or mean cell haemoglobin (MCH) <27 pg, but (because of an a or d-globin chain variant), it is the with normal iron status, will have the Hb A2 quan- total Hb A2 that is diagnostic. It is therefore analytically tification undertaken to help identify their thalassae- important to measure all the Hb A2 and any other hae- mia status. In some other countries, it is up to the moglobin present (usually Hb A and Hb F and not infre- individual and/or doctor to decide if, and when, to quently a Hb A2 variant haemoglobin) to calculate the undertake such tests. relative proportion of Hb A2. It is equally important not Over the last 30 years, great strides have been only to select a method that can detect and quantify made in the screening programmes for the haemoglo- any Hb A2 variant but also for the analyst to be aware binopathies. Despite a wealth of accumulated data, of the need to incorporate them in any analytical proce- several problems still remain in carrier identification. dure. Some variant haemoglobins will interfere with

The most common issue is the presence of mild or the quantiﬁcation of Hb A2; for instance, Hb C and E moderate microcytosis with normal Hb A2 and Hb F migrate with Hb A2 on cellulose acetate electrophoresis that may be due to iron deﬁciency, a-thalassaemia at alkaline pH whereas Hb Lepore, Hb E and sometimes

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 A.D.STEPHENSETAL. Hb A2 QUANTIFICATION AND INTERPRETATION 3

Hb D may co-elute with Hb A2 on many dedicated mutations’ can in fact be the cause of severe transfu- high-performance liquid chromatography (HPLC) sys- sion-dependent conditions in the homozygous state or tems. Some commercial cation-exchange HPLC systems in combination with other b-thalassaemia defects, optimized for the quantification of Hb AIc have com- with all the potential problems associated with b-thal- pressed the later part of the chromatogram to reduce assaemia major (Old et al., 2008). Some b-thalassae- the cycle time, and these are likely to miss many Hb A2 mia carriers have normal or borderline Hb A2 levels variants. but MCV and MCH within the typical carrier range. Kunkel recorded the presence of a minor haemo- Such carriers may include heterozygotes for some globin band on electrophoresis of adult blood (Kunkel ‘mild mutations’, such as IVS-I-6 (TfiC), double het- & Wallenius, 1955) and subsequently showed that it erozygotes for d- and b-thalassaemia (in cis or in trans), was raised in the parents of people with thalassaemia carriers of the Corfu db-thalassaemia and ecdb-thalas- (Kunkel et al., 1957). In people who have the com- saemia heterozygotes. The latter involve very large mon types of b-thalassaemia trait, the Hb A2 concen- deletions of the b-globin gene cluster, some sparing tration is usually between 4.0% and 6.0% and is the b-globin gene (English, Hispanic) and others rarely outside the range of 3.5–7.0% of the total hae- deleting it (Dutch, Scottish–Irish, Mexican, Irish, moglobin. In people who do not have this condition, Canadian, Croatian). Most of these mutant alleles are the Hb A2 concentration is usually between 2.2% and summarized in Table 1. 3.3% and is rarely outside the range of 2.0–3.5% Family studies and DNA analysis are strongly recom- unless severe iron deficiency anaemia is present mended to differentiate these atypical b-thalassaemia (Figure 1). This means that it is essential that the carriers from a-thalassaemia heterozygotes. Atypical measurement of Hb A2 is both precise and accurate carriers with ‘very mild’ or ‘silent’ b-thalassaemia and as stated earlier, if a Hb A2 variant is present, it is mutations (FBC and Hb A2 levels are borderline/nor- the total Hb A2 that has diagnostic significance. Man- mal) may also include carriers of the triple ual assays were developed in the 1960s and 1970s, a-globin gene rearrangement (Giordano, Bakker-Ver- which can give clinically useful results of Hb A2 levels, wij & Harteveld, 2009). Identification of a ‘silent’ and these were reviewed by ICSH and published as b-thalassaemia carrier is often retrospective in parent(s) recommended methods in 1978 (International Com- of patients with moderate/mild b-thalassaemia inter- mittee for Standardization in Haematology, 1978). media. Generally, if a silent b-thalassaemia mutation is It is becoming apparent that many b-thalassaemia suspected, this is best ascertained through DNA analysis mutations associated with ‘borderline’ Hb A2 levels such as b gene sequencing. Because of their ‘silent’ that are sometimes referred to as ‘mild thalassaemia phenotype, these carriers may escape identification in general population screening programmes. In

Table 1. Some b-thalassaemia mutations known to be associated with a borderline Hb A2 levels (Old et al., 2008)

CAP+1 (AfiC) )101 (CfiT) )92 (CfiT) PolyA ()AT) PolyA (AfiG) PolyA ()AA) PolyA (TfiC) IVS-I-6 (TfiC)

Figure 1. Ranges of Hb A2 values measured in 766 IVS-II-844 (CfiG) individuals from different cohorts including 32 cases +1480 (CfiG) of a triplications. +33 (CfiG)

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 4 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION

specialized laboratories, every suspected case is analy- Table 3. Main pre-analytical subject-related factors sed at the molecular level by DNA techniques. It must that may normalize or decrease Hb A2 levels be emphasized that in multi-ethnic populations like those emerging in countries with high levels of immi- Thalassaemic syndromes ‘Silent’ b-thalassaemia gration, the spectrum of mutations can be much larger alleles Interaction between with the likelihood of gene admixture than in countries d- and b-thalassaemia where thalassaemia is endemic. (d+b-thalassaemia) In all these cases, it is extremely important that db-thalassaemia the Hb A2 assay is both accurate and precise as other- Hb H disease and other wise these cases may be missed until the affected a-thalassaemias people produce a child with transfusion-dependent Other haemoglobinopathies a-chain variants d-chain variants thalassaemia. Complicated thalassaemia genotypes Acquired conditions Erythroleukaemia can result in borderline Hb A2 levels and also cause Severe iron-deﬁciency diagnostic problems that may require detailed DNA anaemia analysis to elucidate as mentioned earlier (Paglietti Sideroblastic anaemia et al., 1985; Galanello et al., 1994; Giambona et al., 2008).

Madan et al., 1998). Other conditions have been docu- PRE-ANALYTICAL CONSIDERATIONS mented as reducing the Hb A2 level and are listed in

In 1993, it was documented that the Hb A2 concentra- Table 3. tion may be raised in HIV during treatment (Galacteros et al., 1993; Routy et al., 1993; Pornprasert SAMPLE COLLECTION AND STORAGE et al., 2009). Other causes of a raised level are thought to be very unusual (Weatherall & Clegg, 2001) but Any anticoagulant can be used although it is common are listed in Table 2. Severe iron deﬁciency anaemia to use K2EDTA, which is the anticoagulant used for can reduce the Hb A2 level, (Wasi, Disthasongchan & the analysis of blood counts. Ideally the analysis Na-Nakorn, 1968; Kattamis et al., 1972; Cartei & Dini, should be undertaken as soon as possible after collec- 1975; Cartei et al., 1976; Steinberg, 1993) and in some tion but storage of a sealed whole blood sample at reports, this has been shown to interfere with the 4 C for 2–3 weeks is acceptable for Hb A2 as there diagnosis of beta-thalassaemia trait while others have will be minimal oxidation at 4 C during that time stated that the effect is not severe enough to affect (Tietz, 1990). Where there are high ambient tempera- such diagnosis (Galanello et al., 1990; Steinberg, 1993; tures, such as in tropical areas, suitable means of transport must be used to prevent deterioration of the haemoglobin because considerable denaturation occurs even after 1 h at 50 C. Sample stability after Table 2. Main pre-analytical subject-related factors freezing at )20 or at )80 C is not well documented that may increase Hb A levels 2 although some manufacturers claim an overall stabil- Thalassaemic syndromes Heterozygous b-thalassaemia ity of 1 month at )20 C and 3 months at )80 C. Other Artefact in the presence When freezing samples, it is essential that the blood haemoglobinopathies of Hb S samples are frozen as quickly as possible because slow Some haemoglobin variants freezing of proteins promotes denaturation; one satis- with thalassaemic phenotype factory way is to freeze drops of whole blood or Acquired conditions Hypertrophic osteoarthropathy Megaloblastic anaemia haemolysate aliquots in liquid nitrogen. Thawing and Pseudoxanthoma elasticum refreezing of blood samples are discouraged. A par- Hyperthyroidism tially coagulated blood sample should not be accepted Treatment-related Antiretroviral therapy in for Hb A2 quantiﬁcation by automated equipment as situations patients with HIV it may contaminate or block the tubing or column.

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 A.D.STEPHENSETAL. Hb A2 QUANTIFICATION AND INTERPRETATION 5

commonest causes of a falsely raised Hb A is the EXPRESSION OF RESULTS 2 presence of ‘aged’ Hb S or Hb SId, which frequently

To be clinically useful, the results of the Hb A2 quan- elutes in the same position as Hb A2. The commonest tiﬁcation have to be expressed as a percentage of the cause of a reduced Hb A2 is the presence of an total haemoglobin (which will usually include the a-chain variant that will produce a Hb A2 variant, integration of the Hb A, Hb F, Hb A2 and any glycated which may not be recognized or quantitated. Another or aged adducts). However, in case of elevated base- cause of a Hb A2 variant is a d-chain variant, which lines, one must be careful not to integrate dispropor- again may not be recognized or quantitated. Finally, tional areas (see more under peak quantiﬁcation). the presence of d-thalassaemia will reduce the Hb A2 even in the presence of db-thalassaemia. The implications in interpreting the results in these cases are REFERENCE INTERVALS summarized in Figure 2.

The relative amount of Hb A2 in adult healthy subjects, who are not carriers of b-thalassaemia, is usually PEAK QUANTIFICATION between 2.0% and 3.3% (Weatherall & Clegg, 2001), and some slightly different numbers can be found in With automated systems such as HPLC, capillary zone other contexts, depending on the measurement proce- electrophoresis (CZE) and capillary isoelectric focus- dures and on differences in local populations. As an sing (cIEF), the accuracy of peak quantiﬁcation is example, reference intervals were between 1.9% and dependent on both peak resolution and the method of 3.1% in a study on a Sardinian population (Galanello peak integration. Poor peak quantiﬁcation will result et al., 1994). Other reference laboratories exclude from in poor accuracy but may still result in good precision their normal cohort iron depleted and alpha-thalassae- that may mislead analysts into believing that the mia individuals and consider the variability among dif- accuracy is also good. It is important that the sample ferent instruments, use normal reference intervals loading is within the manufacturer’s recommended between 2.5% and 3.5% (Van Delft et al., 2009) see range as not only must the column not be overloaded Figure 1. but it is essential that the optical absorbance must be kept within the linear range because, if too much or too little sample is applied, any peaks may be under- NEWBORN b-THALASSAEMIA CARRIERS or overestimated leading to the wrong relative estima-

Because of the reduced expression of the d-globin tion of small peaks such as the Hb A2. As mentioned gene with respect to the b gene, the Hb A2 levels at earlier, it is equally important that the baseline used birth are very much lower than in the adult life and for integration is correct as only those parts of the are not measurable by standard methods, but normal- peak that are above the baseline will be quantitated, ize by 12 months of age. In b-thalassaemia carriers, and this may lead to parts of a peak being excluded

Hb A2 may increase with respect to normal subjects from the integration if the baseline is incorrectly already by 3 months after birth (Mosca et al., 2009) assigned. As a general rule, with the equipment used but does not reach levels found in adults until at least for quantitating Hb A2, the baseline should be straight 12 months (Weatherall & Clegg, 2001). Alternatively, and close to the horizontal. Ideally, the baseline the level of Hb A has been used at birth to predict should be obtained by running a blank sample twice with an acceptable probability, the carrier status of (so that there is no haemoglobin present) at the b-thalassaemia in the newborn and conﬁrmed by beginning of each batch and recording the chromato- DNA analysis (Mantikou et al., 2009). gram produced with the second blank (to avoid any carry over), and this should be used as the baseline PITFALLS for that batch of samples. If a haemoglobin peak (such as Hb H) elutes with the void volume and if it occurs As mentioned in ‘analytical considerations’ and sum- with some systems, the integration excludes the ‘void marized in Tables 2 and 3, a number of conditions peak’, then in the presence of Hb H, the total amount may either elevate or reduce the Hb A2. One of the of haemoglobin will be underestimated leading to an

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 6 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION

Figure 2. a variants and d-globin gene defects (d variants and d-thalassaemias) and their implications interpreting the Hb A2 levels in b-thalassaemia diagnosis.

overestimation of the Hb A2 peak. Accurate quantifi- b-thalassaemia trait. Nevertheless, precision is impor- cation of a peak also requires that that peak is com- tant, and the level obtainable with manual tech- pletely separated from neighbouring peaks, and this is niques is ±0.1% in the final answer, which is especially important for small peaks such as Hb A2. equivalent to an SD of 0.05% (leading to a CV of 2% For all these reasons, it is extremely important that in the normal range and 1% in the thalassaemia the chromatogram, or electropherogram, is inspected range). A similar precision level should be obtainable carefully before the results are authorized. As Hb A2 is in the modern automated HPLC and capillary electro- reported as a ratio of Hb A2 to the total haemoglobin, phoretic equipment. Accuracy is harder to establish as long as there is an appropriate baseline associated and monitor but comparison with the WHO Standard with good peak integration, area calibrators should Reference Material (see Conclusions below) and/or not be necessary as they are a poor substitute for comparison with National or Regional quality assess- inadequate chromatography or peak quantification. ment schemes can be helpful. Investigations are in However, standards and controls remain essential to progress to use mass spectrometry to assess candidate verify that the equipment is working satisfactorily. standards. Concerning the relative total error, an Although for HPLC and cIEF it is the peak areas that acceptable limit of 7.0% has been recently proposed are related to the amount of haemoglobin present, it (Mosca et al., 2009), to correctly classify a subject should be noted that in capillary zone electrophoresis, with a true Hb A2 value of 3.6% besides other vari- the peak areas should not be used for quantification ables, such as MCV and MCH. Indeed, in this case, owing to the different migration velocities through the measurement error should not exceed 0.25% the detector, instead the ‘spatial areas’ (integrated (relative total error 7.0%) to avoid the possibility of area divided by the migration time) should be used misclassifying an individual on the basis of a Hb A2 (Huang, Coleman & Zare, 1989; Hempe & Craver, measurement as a putative b-thalassaemia carrier (Hb

1999). A2 >3.8%) or as a non-b-thalassaemia subject (Hb A2 <3.3%). PRECISION, ACCURACY AND DIAGNOSIS ANALYTICAL TECHNIQUES As stated earlier, both precision and accuracy are important because a raised level of Hb A2 is a As stated earlier, the analytical principle consists of diagnostic marker for beta-thalassaemia trait. How- separation and quantiﬁcation of Hb A2 (and any Hb ever, measurement of the Hb A2 level alone cannot A2 variant) in relation to all other haemoglobins pres- absolutely conﬁrm or exclude the carrier state, ent. The haemoglobin fractions can be separated by because there may be very little numerical difference electrophoresis and elution, by anion-exchange micro- between the Hb A2 level in people with and without column chromatography, by cation-exchange HPLC,

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 A.D.STEPHENSETAL. Hb A2 QUANTIFICATION AND INTERPRETATION 7

by CZE or by cIEF. Quantification is usually under- Table 5. Some variant haemoglobins that elute with, or taken by spectrophotometric detection of the sepa- close to, Hb A2 on HPLC rated bands or peaks at 415 nm, which is the absorption maximum for oxyhaemoglobin. It is Hb Abington* important to ensure that bilirubin does not interfere Hb Abruzzo Hb Akron with this analysis because it also absorbs at this Hb Boras wavelength. Packed, washed red cells are used to Hb Bethesda* prepare the haemolysate for electrophoresis and for Hb Chandigarth microcolumn chromatography, the two methods rec- Hb Deer Lodge ommended by the ICSH in 1978 which will remove Hb D Iran* bilirubin. When HPLC is used, bilirubin elutes very Hb Denver* Hb D-Ouled Rabah early with the void peak, and therefore, some Hb E instruments do not include this area in the integra- Hb Ethiopia* tion programme. Hb Barts and Hb H also elute with Hb Fort Worth the void peak. With capillary electrophoresis, the Hb G Copenhagen order of migration is similar to alkaline electropho- Hb G Coushatta* Hb G Ferrara resis. In some new CZE instruments, the resolution Hb G Galveston has been improved so that Hb A2 is resolved from Hb G Honolulu* Hb E. This is also true for some HPLC systems pro- Hb G Taipei vided by some manufacturers. Some analytical fac- Hb Hoshida Hb Hamadan tors that affect Hb A2 quantification are given in Table 4, and some of the haemoglobin variants that Hb Jeddah Hb Kenya elute with, or close to Hb A2, are given in Tables 5 Hb Korle Bu* and 6. The methodological details for electrophoresis Hb Lepore Baltimore and for microcolumn chromatography are given in Hb Lepore Boston the previous ICSH recommendations (International Hb Lepore Hollandia Committee for Standardization in Haematology, Hb Loves Park* Hb M Saskatoon 1978) both of which can give reliable results with Hb Muravera experienced analysts. Hb Nebraska With the huge increase in throughput of Hb A2 Hb Ocho Rios measurements in hospital laboratories because of both Hb Osu Christiansborg* Hb Paddington Hb Rocky Mountain Hb San Bruno* Table 4. Main analytical factors affecting the Hb A2 Hb Santa Juana* quantification Hb SId (the aged adduct of Hb S due to glutathione) The presence of an a or d variant haemoglobin Hb Spanish Town A haemoglobin variant eluting, or migrating, with Hb A 2 Hb Toulon Too large, or too small a sample application Hb Tubingen Incorrect buffer or column temperature* Hb Zurich Inadequate separation of the Hb A2 from Hb A (or any other haemoglobin) Inappropriate baseline† HPLC, high-performance liquid chromatography. Inappropriate integration† *These haemoglobin variants are sometimes labelled Hb A2 and sometimes unknown depending on the instru- CZE, capillary isoelectric focussing; cIEF, capillary isoelec- ment and slight variations in retention times. Even when tric focussing; HPLC, high-performance liquid chroma- the variant does not precisely co-elute with Hb A2,it tography. may still interfere with the Hb A2 measurement and so a *HPLC, CZE and cIEF only. different method of measuring the Hb A2 should be used †HPLC and manual microcolumn chromatography. in this situation.

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 8 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION

haemoglobins such as Hb S, C, D-Punjab, E and Table 6. Some variant haemoglobins that migrate with, O-Arab. However, in spite of these analytical changes, or close to, Hb A2 on capillary zone electrophoresis there have not been any recommendations as to how Hb Chad such equipment should be assessed and selected by Hb E-Saskatoon service, referral or research laboratories. In 1988, the Hb O-Arab WHO agreed on the value of the first International Hb C Harlem* Reference Material for Hb A2 (World Health Organiza- *Insufficient separation for accurate quantification. tion Expert Committee on Biological Standardization, 1994). This is a stabilized freeze-dried preparation (89/666) prepared from blood obtained from the the mandatory and the voluntary screening mother of a patient with beta-thalassaemia major and programmes, automated HPLC or capillary zone elec- is held at National Institute of Biological Standards trophoresis has become the main tool to quantify Hb and Controls (NIBSC) South Mimms, Herts, UK. The

A2 in most laboratories in Europe and North America value assigned to this preparation was obtained by an and are being increasingly used in Asia and Africa, international collaborative study using the methods and capillary isoelectric focussing is also being intro- recommended by ICSH in 1978 and also by the HPLC duced. However, dedicated instruments are available equipment that was commercially available at that from different companies and these approach the ana- time. As it is now 20 years since this international ref- lytical issues in different ways (Van Delft et al., 2009). erence material was prepared and validated, another Some of these automated HPLC instruments are pri- preparation is being developed (Paleari et al., 2010) marily designed to quantify Hb AIc but are also used that will be assayed by mass spectrometry and by the to quantify Hb A2. When these instruments are used original ICSH recommended methods and by HPLC to measure Hb A2, it is essential that the columns and and CZE as well. buffers are optimized for the Hb A2 measurement. Automated instruments are also used as a ﬁrst-line HISTORICAL PERSPECTIVE OF screen for variant haemoglobins such as Hb S, C, D- ELECTROPHORESIS AND COLUMN Punjab, O-Arab and E, and if these haemoglobins are CHROMATOGRAPHY NOMENCLATURE present, they may interfere with the measurement of the Hb A2. However, this fact should not represent a One problem in interpreting and reporting HPLC diagnostic problem because the product of two beta results is the different nomenclature used for elec- genes (Hb A and Hb S, C, E, O or another beta chain trophoretic and chromatographic (HPLC) analysis. variant) conﬁrms the activity of both beta genes and Electrophoresis has been used to investigate human thus the absence of b-thalassaemia and makes the red cell proteins since the 1940s (Stern, Reiner & separation and measurement of the Hb A2 fraction for Silber, 1945; Pauling, Itano et al., 1949; Huisman & b-thalassaemia diagnostics obsolete in these cases Dozy, 1962; Huisman, 1986). Initially, the main (Van Delft et al., 2009). Quality assessment pro- adult haemoglobin band was called Haemoglobin A grammes have shown considerable variation in the (Hb A) and the fetal haemoglobin band was called

Hb A2 levels reported by HPLC analysis from aliquots Haemoglobin F (Hb F). Sickle haemoglobin was dis- of the same blood sample (Batterbee et al., 2010). covered in 1948 (Pauling, Itano et al., 1949) and Interlaboratory variation among users of three HPLC was initially called Haemoglobin B, but this was systems produced by the same manufacturer has been soon changed to Haemoglobin S. Other haemoglobin reported between 6.0% and 9.6% (expressed as CV) variants discovered soon afterwards used the suc- (Paleari et al., 2007). To minimize inter- and intralab- ceeding letters of the alphabet Hb C (Itano & Neel, oratory variation, it is important to use the internal 1950; Hunt & Ingram, 1960); Hb D (Itano, 1951; controls provided by the manufacturers correctly. Baglioni, 1962); and Hb E (Chernoff, Minnich & Automated capillary electrophoresis is becoming avail- Chongchareonsuk, 1954; Hunt & Ingram, 1961), and able and is claimed to be equally satisfactory for mea- later, when most of the letters of the alphabet had suring Hb A2 and also for screening of variant been used up, it was decided to use the place name

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 A.D.STEPHENSETAL. Hb A2 QUANTIFICATION AND INTERPRETATION 9

of the geographical location of the propositus or of POSITIVE IDENTIFICATION OF THE the laboratory describing the new haemoglobin HAEMOGLOBIN FRACTIONS BY PRESENT- (Neel, 1953; Gerald & Ingram, 1961). As stated ear- DAY (2011) COMMERCIALLY AVAILABLE lier, Kunkel (Kunkel & Wallenius, 1955) recognized AUTOMATED HPLC TECHNOLOGY the additional presence of a minor haemoglobin and soon after they showed that it was raised in carriers Knowledge of the behaviour and the chromatographic of thalassaemia and designated it Hb A2 (Kunkel elution patterns is important because it can affect the et al., 1957; Lehmann, 1957). Soon after, they putative identiﬁcation of peaks, their retention times noticed a second minor haemoglobin band (Kunkel and accurate quantiﬁcation; for instance, in a non-

& Bearn, 1957) that was designated Hb A3 by the transfused homozygote for Hb S or a compound HbS/b International Society of Hematology (Lehmann, zero thalassaemia (a person with sickle cell disease),

1957). there is no Hb A but peak(s) due to Hb SIa,b,c often

In the late 1950s and 1960s, column chromatog- elute in a position similar to Hb Ao and often get raphy was used to investigate the heterogeneity of labelled as Hb Ao, hence changing the apparent geno- normal haemoglobin (Hb A and Hb F) and of some type from homozygous Hb S to Hb S/b plus thalassae- variant haemoglobins (such as Hb S and Hb C). mia. Also when Hb S is present, the Hb SId peak often

When these analyses were undertaken, a different elutes in the same place as the Hb A2 peak leading to a nomenclature was used that related to the order of falsely ‘raised’ concentration of Hb A2, incorrectly sug- elution of the haemoglobin off a cationic column gesting that the patient has also inherited b-thalassae- eluted with a buffer of increasing pH and/or ionic mia. Electrophoresis on cellulose acetate at alkaline pH strength. The first (small) peak eluting was called AI will quickly elucidate the presence or absence of Hb A. and the second (main) peak AII and a third (small) On the other hand, the ‘apparent’ Hb Ao peak may in peak AIII. It was soon shown that Kunkel’s Hb A2 fact be a true finding in which case, and in the pres- was similar to the peak labelled AIII and that Kun- ence of microcytosis, the diagnosis could indeed be Hb kel’s Hb A3 was similar to AI (Schnek & Schroeder, S/b plus thalassaemia so long as no a-thalassaemia is 1961). Fetal haemoglobin was also found to consist responsible for the microcytosis. Therefore, although of more than one peak on column chromatography, accurate identification and quantification of Hb A and a small peak: FI and a main peak: FII. It was soon A2 have paramount importance in reaching the correct found that Hb AI could be subdivided into four diagnosis of a large spectrum of Hb S b plus thalassae- smaller peaks called Hb AIa,b,c and d. It was found mia cases, ultimately it will be the DNA analysis that that Hb AIa,b,c were all glycated (Bunn et al., 1976), will precisely define the genotype. whereas Hb AId was formed as an ageing derivative Accurate quantification of Hb S levels in heterozyg- owing to the disulphide linkage of glutathione with otes also has important implications in co-inheritance the b93 cysteine residue (Huisman & Dozy, 1962). of a-thalassaemia (Higgs et al., 1984). In the presence of

As there might be confusion between Hb AII and four normally expressed a genes (aa/aa), the expected

Hb A2,HbAII was renamed Hb Ao; the total Hb A Hb S level will approximately be 38%. In cases with a is therefore the sum of Hb Ao,HbAIa,b,c and Hb plus thalassaemia heterozygosity (-a/aa), the Hb S level

AId. The peak FII was similarly renamed Hb Fo. The, will generally drop to approximately 35% or less. In usually small, Hb FI peak has been shown to be cases with only two functional a-globin genes in CIS acetylated Hb FII; and because it is only a small pro- (–/aa)orTRANS (-a/-a), the Hb S level will drop further portion of the total Hb F, it is only readily seen down to 25%, while in Hb H disease with only one where there is a large amount of Hb F such as in functional a-globin gene left, Hb S may drop further to samples of umbilical cord blood or blood samples around 16%. Such correlations demonstrate the power obtained in the neonatal period. Variant haemoglo- of quantification and its usefulness in reaching a puta- bins, such as Hb S, also subdivide in a similar man- tive diagnosis to be confirmed at the molecular level. ner to Haemoglobin A because of adduct formation DNA analysis will confirm the nature of any hae- and so the total Hb S is the sum of Hb So together moglobin variant, and mass spectrometry has been with Hb SIa,b,c and Hb SId. developed sufficiently to give accurate, detailed

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 10 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION

analysis of haemoglobin even if only small samples C, D-Punjab, E or O-Arab) can be separated and (50 lL) of blood are available (Wild et al., 2001; Wild, quantitated in the same run. Green & Stephens, 2004). • The equipment can usually be operated by relatively junior staff but the interpretation and QC require considerable experience. HIGH-PERFORMANCE LIQUID However, dedicated automated HPLC may also have CHROMATOGRAPHY some disadvantages compared with the manual meth- Several companies now produce HPLC systems. These are ods. First, as stated earlier, the column/buffer combina- all similar, in that all utilize a weak cation-exchange col- tion is sometimes optimized for Hb AIc by compressing umn in their instruments. A sample of a RBC lysate is the ﬁnal part of the chromatogram, and this may mean injected into the system; haemoglobin molecules will that the Hb A2 peak does not separate adequately from adsorb onto the column as they are charged molecules in the Hb Ao peak and also that some Hb A2 variant peaks the buffer system used. An eluting buffer is then injected are missed. Second, it requires analysts with consider- intothesystem;thecompositionofthissolutiongradually able experience of column chromatography and inter- changes by increasing the ionic strength. The pretation of the chromatograms. Some laboratories haemoglobin fractions (normal haemoglobins and any may prefer to prepare their own columns and buffers haemoglobin variant) will elute off of the column when and in that situation, it is their responsibility to ensure theionicstrengthoftheelutingsolutionisgreaterthanthe adequate separation, labelling and integration of the attraction to the column. Different haemoglobin variants various haemoglobin peaks. will have different overall charges because of the amino acid substitution that is present. Thus, the time at which WHEN INTERPRETING HPLC thehaemoglobinmoleculeelutesoffofthecolumn(reten- CHROMATOGRAMS CHECK tion time) is characteristic and reproducible, but not unique, for each haemoglobin variant. As the haemoglo- • That controls elute with the retention time (in the bin fractions elute off the column, they pass through a ‘window’) expected. If they do not do so, it is likely detection system that utilizes absorbance readings at 415 that the buffer/column temperature is incorrect, and 690 nm. The percentage of each haemoglobin frac- although other causes are possible such as an incor- tion (Hb A, Hb F, Hb A2, and any variant) is calculated by rect buffer, column or ﬂow rate, or dirty tubing or an summing the area under each peak in the chromatogram. exhausted column.

Glycosylated haemoglobins (such as Hb AIc) and oxidized • That the peaks are symmetrical. If they are asym- methaemoglobinelutefromthecolumnasseparatepeaks metrical, a variant may be present or the tubing or distinctfromandbeforeHbA.Fastvariants,suchasHbHor column may be contaminated or in a bad condition. HbBartsmaynotbequantifiedastheyusuallyeluteoffthe • Check that all the major and minor peaks expected column before the instrument begins to integrate with are present and check to see whether any unex- many systems designed for adult samples but are usually pected major or minor peaks are present. quantitated in systems designed for neonatal samples • If accurate quantification is required, check that the (VanDelftet al.,2009). baseline is flat for the whole of the chromatogram Automated HPLC has four advantages over the and that the peak is completely separated from manual methods. other peaks and reaches the baseline on both sides.

• Many samples can be loaded at a time, and if necessary, their analysis can continue overnight in the CAPILLARY ZONE ELECTROPHORESIS absence of the analyst. • A much smaller sample size is required, usually of This method utilizes a thin capillary made of fused sil- the order of 50 lL compared with 1–2 mL for the ica with an outer coating of polyimide, usually with manual methods. an inner diameter of 50 or 75 lm. Because the capil-

• Not only Hb A2 but also Hb F and any other com- lary has a very large surface-to-volume ratio, it is mon haemoglobin variants present (such as Hb S, excellent at dissipating the heat generated by the

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION 11

applied voltage. Thus, very large voltages (10–30 kV) generated; the percentage of each haemoglobin frac- can be used, and because of these high voltages, the tion (Hb A, Hb F, Hb A2 and any haemoglobin run times are significantly shortened and the resolu- variants) is calculated. tion increased. The inner surface of the capillary tube has a negative charge because of the bare silica. When ELECTROPHORESIS AND ELUTION an electric field is applied, the buffer solution within the capillary generates an electro-endosmotic flow This gives satisfactory results in experienced hands if (EOF) that moves towards the cathode. A sample of undertaken as stated in the 1978 Recommendations haemolysate is injected into the system, and the elec- (International Committee for Standardization in Hae- tric current applied causes separation of individual matology, 1978). haemoglobins because of differences in overall charges. Haemoglobin variants, if present, may sepa- MICROCOLUMN CHROMATOGRAPHY rate because of a charge difference resulting from the amino acid substitution. However, regardless of the This gives satisfactory results in experienced hands if overall charge of each haemoglobin fraction, the EOF undertaken as stated in the 1978 Recommendations is still stronger than any attraction to either pole, and (International Committee for Standardization in Hae- all haemoglobin fractions will move towards the cath- matology, 1978). ode. All the haemoglobin molecules move past a detector that measures the absorbance at 415 nm. An CONCLUSIONS electropherogram (similar to a chromatogram) is thus generated; the percentage of each haemoglobin frac- • It is essential to be able to measure the proportion of tion (Hb A, Hb F, Hb A2 and any haemoglobin vari- Hb A2 in a red cell haemolysate in order to make, or ants) is calculated (but see the paragraph above on exclude, the diagnosis of b-thalassaemia trait. peak integration as ‘spatial areas’ have to be used for • As there may be only a small numerical difference quantification because the fractions pass the detector in Hb A2 levels between people who have b-thalas- at different velocities). For haemoglobinopathy work, saemia trait and those who do not, precision and capillary zone electrophoresis has the same advantages accuracy are both important. However, excluding as HPLC over manual methods and some equipment beta-thalassaemia, because the Hb A2 level is 0.2% and reagents also have an advantage over HPLC in below the cut-off is not recommended, and one that haemoglobin adducts such as glycated XIa,b,c and should always include all haematological parameters the ageing glutathione adduct, XId, do not separate in reaching a conclusion, and in case of doubt, DNA from the main haemoglobin peaks, and this makes analysis should be undertaken. interpretation easier than with HPLC. Some variant • Methods used in clinical laboratories first involve haemoglobins (Table 6) migrate with or very close to separating the Hb A2 from Hb A and any other hae-

Hb A2 and therefore interfere with the quantification moglobin present and then quantitating the differ- of Hb A2. ent fractions present. • Historical methods of analysis (cellulose acetate electrophoresis with elution or microcolumn chro- CAPILLARY ISOELECTRIC FOCUSSING matography) that were discussed in a previous In cIEF, the electro-osmotic flow is minimized by publication (International Committee for Standardi- coating the inside of the capillary with a solution to zation in Haematology, 1978) can produce results neutralize the silanols. The capillary is filled with a that are satisfactory for clinical purposes but are mixture of ampholytes and haemolysate. After appli- time-consuming and need technical expertize. cation of the voltage for a period of time, the haemo- • Automated HPLC with dedicated commercially avail- globins become focussed at their pI value. They are able buffers and columns is increasingly being intro- then mobilized and pass through the detector operat- duced and can give satisfactory results when no ing at 415 nm (Hempe & Craver, 1999). An electro- haemoglobin variant is present. However, when pherogram (similar to a chromatogram) is thus haemoglobin variants are present, experience is

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 12 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION

necessary to interpret the chromatograms. It is essen- • At present, the only internationally recognized stan-

tial that the reagents are optimized for Hb A2 rather dard for Hb A2 is the WHO ﬁrst Reference Material

than Hb AIc and that the baseline is appropriate. (89/666) that is available from NIBSC, Blanche • Automated capillary electrophoresis and automated Lane, South Mimms, Potters Bar, Hertfordshire, capillary isoelectric focussing with commercially EN6 3QG, UK (http://www.nibsc.ac.uk). It is hoped available capillaries and buffers are becoming avail- that in the foreseeable future, a new certiﬁed refer- able and can also produce satisfactory results if no ence material will be available from the European haemoglobin variant is present, but experience is Commission Joint Research Centre, Institute for necessary to interpret the electropherograms. Reference Materials and Measurements.

REFERENCES heterozygous for beta-thalassemia. Haemato- Journal of Clinical and Laboratory Medicine logica 75, 319–322. 60, 302–318. Baglioni C. (1962) Abnormal human haemo- Galanello R., Barella S., Ideo A., Gasperini Hunt J.A. & Ingram V.M. (1960) Abnormal globins. VII. Chemical studies on haemoglo- D., Rosatelli C., Paderi L., Paglietti E., human haemoglobins. IV. The chemical dif- bin D. Biochimica et Biophysica Acta 59, Sollaino C., Perseu L., Loi D. & Cao A. ference between normal human haemoglobin 437–449. (1994) Genotype of subjects with border- and haemoglobin C. Biochimica et Biophysica Batterbee H., De La Salle B., Mctaggart P., Dore line hemoglobin A2 levels: implication for Acta 42, 409–421. C., Wild B. & Hyde K. 2010. Evaluation of beta-thalassemia carrier screening. American Hunt J.A. & Ingram V.M. (1961) Abnormal

UKNEQAS(H) Hb A2 and related performance Journal of Hematology 46, 79–81. human haemoglobins. VI. The chemical dif- data [Online]. Available at: http:// Gerald P.S. & Ingram V.M. 1961. Recommenda- ference between haemoglobins A and E. Bio- www.ukneqash.org (accessed 1 February tions for the nomenclature of hemoglobins. chimica et Biophysica Acta 49, 520–536. 2011). The Journal of Biological Chemistry 236, International Committee for Standardization in Baysal E. (2001) Hemoglobinopathies in the 2155–2156. Haematology (1978) Recommendations for United Arab Emirates. Hemoglobin 25, 247– Giambona A., Passarello C., Vinciguerra M., Li selected methods for quantitative estimation of 253. Muli R., Teresi P., Anza M., Ruggeri G., Hb A2 and for Hb A2 reference preparation. Bouva M.J., Harteveld C.L., Van Delft P. & Renda D. & Maggio A. (2008) Signiﬁcance British Journal of Haematology 38, 573–578. Giordano P.C. (2006) Known and new delta of borderline hemoglobin A2 values in an Itano H.A. (1951) A third abnormal hemoglobin globin gene mutations and their diagnostic Italian population with a high prevalence of associated with hereditary hemolytic anemia. signiﬁcance. Haematologica 91, 129–132. beta-thalassemia. Haematologica 93, 1380– Proceedings of the National Academy of Sci- Bunn H.F., Haney D.N., Kamin S., Gabbay K.H. 1384. ences of the United States of America 37, & Gallop P.M. (1976) The biosynthesis of Giordano P.C., Bakker-Verwij M. & Harteveld 775–784. human hemoglobin A1c. Slow glycosylation C.L. (2009) Frequency of alpha-globin gene Itano H.A. & Neel J.V. (1950) A new inherited of hemoglobin in vivo. Journal of Clinical triplications and their interaction with beta- abnormality of human hemoglobin. Proceed- Investigation 57, 1652–1659. thalassemia mutations. Hemoglobin 33, 124– ings of the National Academy of Sciences of Cartei G. & Dini E. (1975) Relationship 131. the United States of America 36, 613–617.

between Hb and HbA2 concentrations in Hempe J.M. & Craver R.D. 1999. Laboratory Kattamis C., Lagos P., Metaxotou-Mavromati A. untreated and Fe-treated iron deﬁciency diagnosis of structural hemoglobinopathies & Matsaniotis N. (1972) Serum iron and anaemia. Biomedicine 23, 85–87. and Thalssemias by capillary isoelectric focus- unsaturated iron-binding capacity in the - Cartei G., Chisesi T., Cazzavillan M., Battista R., ing. In: Clinical Applications of Capillary thalassaemia trait: their relation to the levels Barbui T. & Dini E. (1976) Relationship Electrophoresis (ed. S.M. Palfreys), 81–98. of haemoglobins A, A 2, and F. Journal of

between Hb and HbA2 concentrations in Humana Press, Totowa, New Jersey. Medical Genetics 9, 154–159. beta-thalassemia trait and effect of iron deﬁ- Higgs D.R., Clegg J.B., Weatherall D.J., Serjeant Kunkel H.G. & Bearn A.G. (1957) Minor hemo- ciency anaemia. Biomedicine 25, 282–284. B.E. & Serjeant G.R. (1984) Interaction of the globin components of normal human blood. Chernoff A.I., Minnich V. & Chongchareonsuk alpha alpha alpha globin gene haplotype and Federation Proceedings 16, 760–762. S. (1954) Hemoglobin E, a hereditary abnor- sickle haemoglobin. British Journal Haema- Kunkel H.G. & Wallenius G. (1955) New hemo- mality of human hemoglobin. Science 120, tology 57, 671–678. globin in normal adult blood. Science, New 605–606. Huang X., Coleman W.F. & Zare R.N. (1989) Series 122, 288. Galacteros F., Amaudric F., Prehu C., Feingold Analysis of factors causing peak broadening Kunkel H.G., Ceppellinini R., Mu¨ ller-Eberhard N., Doucet-Populaire F., Sobel A. & Rosa J. in capillary zone electrophoresis. Journal of U. & Wolf J. (1957) Observations on the (1993) Acquired unbalanced hemoglobin Chromatography 480, 95–110. minor basic hemoglobin component in the chain synthesis during HIV infection. Comp- Huisman T.H.J. 1986. The Hemoglobinopathies. blood of normal individuals and patients with tes Rendus de l’Academie des Sciences. Serie Methods in Hematology. Churchill Living- thalassaemia. Journal of Clinical Investigation III, Sciences de la Vie 316, 437–440. stone, Edinburgh. 36, 1615–1625. Galanello R., Turco M.P., Barella S., Giagu N., Huisman T.H.J. & Dozy A.M. (1962) Studies on Lehmann H. (1957) The hemoglobinopathies Dessi C., Cornacchia G. & Cao A. (1990) Iron the heterogeneity of hemoglobin. V. Binding (Report of the ISH Meeting in Boston). Blood stores and iron deﬁciency anemia in children of hemoglobin with oxidised glutathione. 12, 90–92.

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13 A. D. STEPHENS ET AL. Hb A2 QUANTIFICATION AND INTERPRETATION 13

Madan N., Sikka M., Sharma S. & Rusia U. commercial HPLC systems. Clinical Chemistry Journal of the Medical Sciences 305, 312–

(1998) Haematological parameters and HbA2 and Laboratory Medicine 45, 88–92. 313. levels in beta-thalassaemia trait with coinci- Paleari R., Munoz A., Mosca A. & IFCC Work- Stern K., Reiner M. & Silber R. 1945. On the

dent iron deﬁciency. Indian Journal of ing Group on Standardization of HBA2 (WG- electrophoretic pattern of red blood cell pro-

Pathology and Microbiology 41, 309–313. HBA2) (2010) Towards the development of a teins. Journal of Biological Chemistry 161, Mantikou E., Arkesteijn S.G., Beckhoven Van certiﬁed reference material for hemoglobin 731.

J.M., Kerkhoffs J.L., Harteveld C.L. & Giord- A2. Clinical Chemistry and Laboratory Medi- Tietz N., Finley P.R. & Pruden E.L. (1990) Clini- ano P.C. (2009) A brief review on newborn cine 48, 1611–1618. cal Guide to Laboratory Tests, 2nd edn. W.B. screening methods for hemoglobinopathies Pauling L., Itano H.A., Singer S.J. & Wells I.C. Saunders Company, Philadelphia. and preliminary results selecting beta thalas- (1949) Sickle cell anemia, a molecular dis- Van Delft P., Lenters E., Bakker-Verweij M., De semia carriers at birth by quantitative estima- ease. Science 109, 443. Korte M., Baylan U., Harteveld C.L. & Giord- tion of the HbA fraction. Clinical Phylipsen M., Gallivan M.V., Arkesteeijn S.G., ano P.C. (2009) Evaluating ﬁve dedicated Biochemistry 42, 1780–1785. Harteveld C.L. & Giordano P.C. (2011) Occur- automatic devices for haemoglobinopathy Mosca A., Paleari R., Ivaldi G., Galanello R. & rence of common and rare d-globin gene diagnostics in multi-ethnic populations. Inter- Giordano P.C. (2009) The role of haemoglo- defects in two multiethnic populations:thir- national Journal of Laboratory Hematology

bin A2 testing in the diagnosis of thalassae- teen new mutationsand the signiﬁcance of 31, 484–495. mias and related haemoglobinopathies. d-globin gene defects in b-thalassaemia diag- Wasi P., Disthasongchan P. & Na-Nakorn S. Journal of Clinical Pathology 62, 13–17. nostics. International Journal of Laboratory (1968) The effect of iron deﬁciency on

Neel J. (1953) Statement concerning a system Hematology 33, 85–91. the levels of hemoglobins A2 and E. Journal of nomenclature for the varieties of human Pornprasert S., Sukunthamala K., Leechanachai of Laboratory and Clinical Medicine 71, 85– hemoglobin. Blood 8, 386–387. P. & Sanguansermsri T. (2009) Increased Hb 91.

Old J., Traeger-Synodinos J., Christou S., Borg A2 values in an HIV-1-infected patient receiv- Weatherall D.J. & Clegg J.B. (2001) The Thalas- J., Molyneux A., Henderson S. & Fattoum S. ing antiretroviral drugs: a pitfall for thalasse- saemia Syndrome. Blackwell Scientiﬁc,

2008. Normal Hb A2 b-thalassaemia in Eur- mia antenatal diagnosis. Hemoglobin 33, Oxford. ope – characterisation of eight mutations, 158–161. Wild B.J., Green B.N. & Stephens A.D. (2004) including a novel Poly A mutation (-AA). Routy J.P., Monte M., Beaulieu R., Toma E., The potential of electrospray ionisation mass 10th International Conference on Thalassae- St-Pierre L. & Dumont M. (1993) Increase of spectrometry for the diagnosis of hemoglobin mia & Haemoglobinopathies: abstracts pub- hemoglobin A2 in human immunodeﬁciency variants found in newborn screening. Blood lished in conference proceedings, Singapore: virus-1-infected patients treated with zidovu- Cells, Molecules and Diseases 33, 308–317. EPI 49, 99. dine. American Journal of Hematology 43, Wild B.J., Green B.N., Cooper E.K., Lalloz Paglietti E., Galanello R., Addis M. & Cao A. 86–90. M.R.A., Erten S., Stephens A.D. & Layton (1985) Genetic counseling and genetic heter- Schnek A.G. & Schroeder W.A. 1961. The rela- D.M. (2001) Rapid identiﬁcation of haemo- ogeneity in the thalassemias. Clinical Genet- tion between the minor components of globin variants by electrospray ionisation ics 28, 1–7. whole normal human adult hemoglobin as mass spectrometry. Blood Cells, Molecules Paleari R., Giambona A., Cannata M., Leto F., isolated by chromatography and starch block and Diseases 27, 691–704. Maggio A., Mosca A. & For the IFCC Work- electrophoresis. Journal of the American World Health Organization Expert Committee

ing Group ‘‘Standardization of HBA2’’. (2007) Chemical Society 83, 1472–1478. on Biological Standardization 1994. Biological

External quality assessment of hemoglobin A2 Steinberg M.H. (1993) Case report: effects of Standardization Forty Fourth Report. WHO measurement: data from an Italian pilot iron deﬁciency and the -88 C–>T mutation Technical Report Series – 848. World Health

study with fresh whole blood samples and on HbA2 levels in beta-thalassemia. American Organization, Geneva.

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2012, 34, 1–13